This proposal investigates the role acoustic prominence, which is one aspect of prosody, plays in language production and comprehension. Traditionally, researchers have characterized acoustic prominence as a means by which a speaker signals the givenness of information: new information is accented and given information is not. In this proposal it is proposed that acoustic prominence is a function of 1) the importance of a word to a conversation and 2) the predictability of the word. Results from preliminary experiments suggest that speakers produce important, less predictable words with more prominence, and that informativeness and predictability play a role in the type of prominence that is chosen by speakers. However, important questions remain. Experiments 1-6 investigate whether the degree of acoustic prominence produced by a speaker is proportional to its importance in a referential communication task and whether listeners are sensitive to fine- grained differences in acoustic prominence using an eye-tracking visual world paradigm. Experiments 7-8 explore whether effects of predictability on acoustic prominence are driven by processes in planning or by processes that facilitate listener comprehension. Experiments 9 and 10 investigate how predictability and informativeness interact to influence pitch accent choice. PUBLIC HEALTH RELEVANCE An important part of successful communication is producing and understanding language with appropriate prosody (e.g. stress, pitch, rhythm, and pausing). Individuals with language deficits that affect the production and comprehension of prosody are likely to have difficulty communicating, and little is known about how prosody facilitates communication in on-line language processing. This proposal investigates the role acoustic prominence, which is one aspect of prosody, plays in language production and comprehension.